-
Notifications
You must be signed in to change notification settings - Fork 843
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support sub-library dependencies (rebased). #5839
Conversation
1ea2757
to
f10bc53
Compare
I have another much larger repo that I am testing this on and it fails because, even though I've built stack with
How do I get stack to go all-in on stack/src/Stack/Types/Config.hs Lines 2039 to 2045 in 55291b3
|
@philderbeast, I think the |
I don't know if this is relevant for your pull request, but I am keen to move Stack on to GHC 9.4.x as soon as a Stackage resolver is available and covers Stack's dependencies. That is because recent changes in MSYS2 can be problematic for Windows users of GHC < 9.4.1, as explained in GHC issue #21111. |
The test of this pull request on resolver: nightly-2022-07-10
install-ghc: false
system-ghc: true
packages:
- simple-smt
- ghc-tcplugins-trace/ghc-tcplugins-trace
- build
- .
extra-deps:
- ghc-corroborate-1.0.0@sha256:60e580741703235dc81232c41b6a7b29b5cf5e337c92ef1745509e8abf22d2b5,3855 |
I'm good with that. I did a little more testing with plugins-for-blobs and found that sub-library dependencies work for
|
I think the CI failures may be something to do with the module name clashes discussed here: https://hackage.haskell.org/package/Cabal-syntax-3.6.0.0. |
Looks like #5691 I believe if you constrain |
Quite so. The command that fails shows that
|
@philderbeast, the |
@philderbeast, do you anticipate returning to this pull request? It seemed that you were close to your goal. |
Yes, I think I'll be able to get to it tomorrow. |
@philderbeast, I would be happy to help with the conflicts with the current |
@mpilgrem thanks for the offer to help with the merge. I've cherry-picked the one commit, 9ddf6f1, that matters for adding sub-library dependencies on branch add/sublib-deps-redo and am starting to test it. We have the integration test |
@philderbeast, the error in the integration test - |
More information about the error in the integration test (extracts, from a local build):
Looking at:
Also:
|
@philderbeast, I've been thinking about the failing integration test. I do not yet have an answer, but it seems to me that the underlying problem is that the I am going to do two minor things in the interim:
|
@mpilgrem if you'd like to try a project with a lot of sublibs then here are stack and cabal commands for working with plugins-for-blobs. At first glance both seem to work and I'm using 39a8253 for the stack build and
The tests pass except for doctests that have come loose. A z3 release is required for the compilation.
During the build of plugins-for-blobs, I dump a lot of output to the console like the following and that is normal (helps debugging the conversion from GHC typechecking constraints into SMT-LIB expressions):
|
@philderbeast, I thought, locally, that I had a solution (to the failing integration test I think the solution lies in a change to |
@mpilgrem I added a |
d5e6c85
to
89c0fee
Compare
I rebased on the current master branch. |
7a1c389
to
9e0a4b7
Compare
457a9d9
to
3cf0241
Compare
@philderbeast, I am making little progress on this. I have, however, moved the parts that I think are sound into Stack's |
I am closing this pull request, as it has been overtaken by #6343, which is close to being merged. |
I brushed the dust off #5659, rebased it and tested it with plugins-for-blobs. Seems to work.